A Review on the Classification of Body Movement Time Series to Support Clinical Decision-Making

Classifying body movement time series has supported decision-making in many medical specialties, such as Neurology and Psychiatry. However, a protocol-guided review of the recent literature on this topic is missing. This work aims to supplement earlier surveys by designing and conducting a review of the literature on classification approaches of body movement time series that have been applied to support clinical decision-making. To do so, we defined a protocol that indicates a research question, a study search strategy, a selection strategy, some information items to be extracted and a synthesis strategy. By applying the survey method, 55 papers published until 2023 were selected and summarized. As a result, it was found that these references follow a basic workflow based on time series acquisition, characterization, classification, and performance evaluation. Information from the 55 publications can be found in a publicly available spreadsheet. This work also indicates additional considerations and future research directions.